Why TUTUR exists
AI chat that truly understands your context. TUTUR combines retrieval, generation, and memory in one platform.
The problem
Most AI chatbots just throw questions at an LLM and return answers. Without context from your data, answers are often irrelevant or hallucinated. Without memory, every conversation starts from zero.
Building proper AI chat requires many components: vector database, embedding pipeline, chunking strategy, prompt engineering, session management, and memory system. Each component has its own complexity.
TUTUR unifies all these components in one ready-to-use platform. RAG for accuracy, memory for context, multi-tenant for scalability.
Four principles
TUTUR's design foundations.
RAG-First
Every answer is grounded in your data. Retrieval is not optional — it's the foundation. Without grounding, AI just guesses.
Memory Matters
Conversation without memory is not a conversation. 4-layer memory system ensures the chatbot remembers what matters.
Provider-Agnostic
Not locked into one AI provider. OpenAI, DeepSeek, Groq — choose the best, switch anytime.
Tenant Isolation
Each tenant's data is fully isolated. One platform, many chatbots, zero data leakage.
ARSAKA Ecosystem
TUTUR is part of the ARSAKA ecosystem — an integrated platform for engineering teams.
PUGUH
ActiveAuthentication, IAM, multi-tenancy, billing, audit. Foundation for all ARSAKA products.
MANTRA
ActiveConstitutional law system for architecture decisions. Immutable records, 3-gate validation, smart retrieval.
TUTUR
ActiveAI chat with RAG, multi-provider, vector search, 4-layer memory, and multi-tenant architecture.
No vendor lock-in
Your data stays yours. Export conversation history, knowledge base, and configuration anytime. Switch AI providers without migration. Self-hosted option available for Enterprise.
"Your data, your AI, your rules."